Many conventional mechanical systems are monitored to determine the health thereof and the vibrations caused thereby. However, there are many instances where it is desirable to make broadband measurements of vibration or other physical quantities at locations where it is extremely difficult to do so. For example, the measurement environment may be a harsh environment in which sensors are unable to operate reliably, an environment in which access is so limited that a sensor cannot be physically mounted at the desired location, or an environment in which access is so limited that a sensor cannot be replaced or repaired in the event of failure.
Some conventional methods of dealing with the above issues have generally involved embedding a piezoelectric or strain sensor at the location to be measured and acknowledging that its lifetime and measurement capabilities are limited by the environment within which it is configured. Other conventional methods of dealing with the above issues have involved a mechanical waveguide transmitting a signal at a single, stationary frequency or a signal for which it is unimportant to detect the correct relative phase and/or amplitude of various frequencies. For example, in ultrasonic non-destructive testing of metal die castings, a simple metal rod can act as a waveguide to isolate an ultrasonic sensor from a hot test location, with one end of the rod at the location to be tested and the other end connected to the sensor. As the casting is heated it expands and as the casting cools it shrinks, which can result in cracks within the casting. Each crack, in turn, can be detected using very high frequency (VHF) energy transmitted through the waveguide. However, when the measurement of the actual amplitude and/or phase characteristics of the signal across a frequency band is important rather than the mere detection of cracks using average VHF energy or other techniques using a fixed frequency signal, a traditional waveguide cannot be used, as the ultrasonic signal reflects within the waveguide, resulting in resonances which distort the signal characteristics, making broadband measurements generally impossible.
This drawback of conventional waveguides is that they often have one or more resonant frequencies that in turn causes large peaks and valleys in the spectrum of any signals transmitted thereby. Resonances are often caused by reflections of a signal from the ends of the waveguide. More particularly, a signal initiated at the sensing end of the waveguide travels down the waveguide and reflects off the non-sensing end, then travels back down the waveguide to the sensing end, is reflected off the sensing end back to the non-sensing end, etc. Considered from the perspective of the frequency domain, if there is a periodic signal at a frequency such that the period of the signal is equal to the time it takes for the signal to travel to the end of the waveguide and back to the sensing end, the reflection and the periodic signal will be in phase and re-enforce each other. This creates a standing wave or resonance in the waveguide. The amplitude of the signal response is very large at these frequencies and is very low at intermediate frequencies where the periodic signal and reflected signal cancel each other out rather than re-enforce. This significantly affects sensor readings, rendering detection of a particular signal generally useless.
Consequently, there is a continuing need for improving signal detection with waveguides to address these and other difficulties with conventional waveguide technology.